Laser Magnetic Resonance Spectroscopy of C10 and Kinetic Studies of the

Reactions of C10 with NO and NO2

YUAN-PERN LEE,* RICHARD M. STIMPFLE,tZ RCBERT A.

DONALD A. JENNINGS," and CARLETON J. HOWARD7 L c

PERRY,** JOHN A. MUCHA,s KENNETH M. EYENSON,"

Aeronomy Laboratory. N O M Environmental Research Laboratory, Boulder, Colorado 80303

Abstract

Far-infrared rotational transitions in ClO[X2IIa, u = 0) have been observed using laser magnetic resonance (LMR) with an o p t i d y pumped spectrometer. Five observed transitions at wavelengths between 444 and 713 pm have been compared with values predicted with spectroscopic constants from the literature. LhfR detection of C10 has been used to study its reactions with NO and NO2 in a discharge flow system under pseudo-fust-order conditions for C10. The measured rate constants are k(C10 + NO) = (7.1 f 1.4) X erp[(270 f 5O)/T] cmVm0le~-s for the temperature range of 202 < T < 393 K k(C10 + NO2 + M) = (28 f 0.6) X 10-33 ex1;;<1090 f sO)/T] cm6/mole& (M = He, 250 < T < 387 K), (3.5 f 0.6) X lo-= exp[(ll80 f 80)/T] (M = 02,250 < T < 416 K), and (2.09 f 0.3) X (M = N2, T = 297 K). All measurements were made a t low pressures, between 0.6 and 6.6 ton. These results are compared with those from other studies.

Introduction

The potential impact of stratospheric ozone depletion by the chlorine catalytic cycle __

t (1) c1+ 03 - e10 + 0 2

c10 + 0 - c1+ 0 2

Net: 0 3 + 0 - 202

-

- (2) 1

NOM-CIRES Postdoctoral Research Associate. Present address: Chemistry De-

t Also affiliated with the Department of Chemistry, University of Colorado, Boulder, CO

* Present address: Division of Applied Sciences, Harvard University, Cambridge, MA

** NOM-NRC Postdoctoral Research Associate, Present address. Combustion Research .

8 NBS-NRC Postdoctoral Research Associate. Present address Bell Laboratories, Murray

O National Bureau of Standards, Time and Frequency Division, Boulder, CO 80303.

partment, National Tsing-Hua University, Hsin-Chu, Taiwan, Republic of China.

80309.

02138.

Facility, Sandia Laboratories, Livermore, CA 94550.

Hill, NJ 07974.

- -

International Journal of Chemical Kinetics, Vol. 14,711-732 (1982) 0 1982 John Wiley & Sons, Inc. CCC 0538-8066/82/060711-22$03.20

c

712 LEE ET’&

has generated considerable interest [l]. To evaluate the effects of chlorine on ozone one must consider the reactions involving the active species C1 and C10. The presence of nitrogen oxides (NO and NO21 in the strato- sphere limit the effectiveness of the chlorine cycle. First, the reaction with NO, (3) C10 + NO -. C1+ NO2 when combined with

(4) NO2 + hv(X < 380 nm) - NO + 0

(5 ) 0 + 0 2 + M -C 03 + M and reaction (1) produces no net destruction of ozone. Second, the reacxion of C10 with NO2,

(6) CIO + NO2 + M -+ C1N03 + M forms ClN03 [2], which is thought to be an unreactive reservoir for chlo- rine.

There have been five studies of the C10 + NO reaction. Clyne and Watson [3] first measured the rate constant K = (1.7 f 0.2) X cm3/ molec-s at room temperature using discharge-flow/mass spectrometry in 1974. Later Zahniser and Kaufman [4] employed the discharge-flow/ resonance fluorescence technique to determine the rate constant relative to that of the C1+ O3 reaction. They reported rate constants at temper- atures ranging from 230 to 295 K, giving a room temperature value slightly higher than the previous value and a negative Arrhenius activation energy of 400 caVmol (1 cal = 4.184 J). Leu and DeMore [5] investigated the temperature dependence (227-415 K) using mass spectrometry and ob- tained slightly lower rate constants and a negative activation energy of about 590 cal/mol. Recently Clyne and MacRobert [6] and Ray and Watson [7] repeated the measurement at room temperature and reported rate constants close to the first study.

The ClO + NO2 + M reaction in the low-pressure range of 1-6 torr (1 ton = 133.3 Pa) has been studied by three groups (8-101 using discharge-flow techniques with various detection methods. Leu et al. [SI and Birks et al. [9] detected C10 by mass spectrometry. This technique was complicated - by effects of ClN03 fragmentation in the mass spectrometer ion source to produce C10. Since the rate constant measurement is based on the pseudo-first-order decay of C10, this effect could lead to an underestimation of the true value of kg. Birks et al. used an empirically derived correction . for this effect that increased the measured value by up to about 17%, while Leu et al. operated under conditions which they stated required no cor- rection. Zahniser et al. [lo] detected C10 indirectly by conversion of C10 to C1 with NO addition, reaction (3), followed by resonance fluorescence detection of the C1 atoms.

.

-

- LMR STUDIES 6F CIO 713

While the various rate constant measurements for both reactions (3) and (6) are in reasonable agreement, some significant discrepancies still exist. For the C10 + NO reaction, the rate constants given by the two temperature dependence studies [4,5] differ by about 40%. For the C10 + NO2 + M reaction the reported rate constants at room temperature differ by about 30% in He and by about 20% in Nz. In addition, the reactit 3

17) C10 + 0 2 + M - OC102 + M may take place at high 0 2 concentrations and could have an effect on the C10 + NO2 + 0 2 reaction. Since no measurements of hj were made in 0 2 ,

it is interesting to investigate reaction (6) in 0 2 as a first step toward the understanding of the importance of reaction (7) and the chemistry of OClOz in the stratosphere. Also it is worthwhile to test a new method which de- tects C10 directly, has high sensitivity, and does not have possible com- plications from the product ClN03 or other species.

Here we report the far-infrared (FIR) laser magnetic resonance (LMR) spectra of C10 at five different laser frequencies. Applying this LMR de- tection technique, we have also investigated the temperature dependence of the rate constants of reactions (3) and (6).

Experimental

An LMR spectrometer w~ used in conjunction with a discharge-flow tube in this study. The basic principles of LMR for spectroscopic as well as kinetic studies have been described previously [11,12]. Briefly, a con- tinuous-wave COz laser with an output power of approximately 25 W was used to pump a FIR CHsOH, CD31, or DCOOH laser which produces about 100 pW. The intracavity absorption was accomplished by Zeeman tuning a rotational transition of the C10 molecule into resonance with the FIR laser frequency. The magnetic field can be varied up to 21 kG. High sensitivity was sbtained by ac modulating the magnetic field and using frequency- locked phase-sensitive detection. It has been found that the peak-to-peak height of the first derivative of the absorption curve is proportional to the radical concentration [ll].

In the kinetic studies, the rat$ constants were measured by observing the pseudo-fmt-order decay of C10 concentrations in a large excess of NO or NO2 as a function of reaction time. The C10 radical was detected at -0.5 kG using a r-polarized 556.9-pm CDBI laser line pumped by the lOP(36) CO2 laser emission. The detection limit for C10 is approximately 2 x 108 molec/cm3 (si:rxl-to-noise ratiq N 1 at 1 s time constant).

The discharge-flow apparatus [12] consists of a 2.54-cm inside diameter, 110-cm long Pyrex flow tube with a 9-mm outside diameter movable inlet. The flow tube was mounted in the LMR spectrometer with the tube exit positioned directly above the intersection of the FIR laser beam and the magnetic field.

- -

.

714 LEE ET~AL.

The total gas flow consisted of the carrier gas (He, 0 2 , or Nz), NO2 or NO, and the gas mixture from the C10 source. For the C10 + NO2 + M reaction rate measurements, the C10 source was part of the movable inlet which allowed variation of the position of the C10 injection into the flow tube. The C10 radical was formed by reacting C1 atoms with excess of 0 3 in the top part of the movable inlet (1) C l + 0 3 - + C 1 0 + 0 2 The C1 was generated by passing a trace of C12 N 0.002 STP cm3/s (STP = 273 K, 1 atom) in helium carrier gas (0.2-0.4 STP cm3/s) through a 20-35-W microwave discharge. Ozone was eluted with helium from a 196 K silica gel trap and diluted with additional helium (0.3-1.0 STP cm3/s) before reacting with C1 atoms. The C10 formation is >99% completed within the first 5-cm length of injector based on the reported value [13] of kl = 1.1 X 10-1’ cm3/molec.s and the ozone concentration of about 3 X 1014 molec/cm3. The distance from the 0 3 inlet to the end of the injector was about 125 cm, long enough for any vibrationally excited C10 formed by reaction (1) to be quenched before it entered the reaction region. The C10 concentration in these experiments was about (1-8) X 1OI2 molec/cm3. A t this concentration C10 does not react significantly with itself in the flow tube [13].

For the C10 + NO2 + M study, NO2 was admitted into the flow tube through a side arm about 80 cm upstream of the LMR detection region. The NO2 concentration in the flow tube ranged from 2 to 20 X 1014 molec/cm3. The NO2 flow rate was measured by observing the rate of the pressure change in a calibrated volume. The effect of NO2 dimerization

- has been taken into account in the NO2 flow rate measurements by applying the correction F N O ~ = J’”oz(l + P / K P ) , where J’”02 is the measured ap- parent NO2 flow rate, H (in torr) is the average pressure in the calibrated volume, and K,, = [NO2]2/[N204] is the equilibrium constant (in torr) for the NO2 dimerization reaction. A t 298 K, KP = 106 torr [14], and the NO2 flow-rate correction is less than 5s’, with P cv 2’torr. NO2 flow measure- ments were always made at low pressure to minimize the correcrion. The effect of NO2 dimerization in the flow tube w e negligible (Cl5 1 a t tem- peratures above 270 K for [NO*] C 2 X 1015 molec/cm3 (KP -N 9.7 torr a t 270 K).1 However, at 250 K KP decreased to 1.18 torr [15], and about a 4% -

decrease in NO2 concentration due to dimerization should be taken into accoynt for [NO21 N 1 X 1015 molec/cm3.

For the C10 + NO reaction, the C10 was formed in the flow tube by re- acting ozone, (3-12) X 1011 molec/cm3, with excess C1 atoms, (8-15) X 1013 atom/cm3, which were produced by a microwave discharge of C12 and He in a sidearm of the flow tube. NO was mixed with -0.5 STP cm3/s of he- Iium and added through the movable injector. The NO concentration in the flow tube ranged from (3 to 30) X 1OI2 molec/cm3.

-

. ’

1 Obtained using thermochemical data from [15].

LMR STUDIES OF C10 715

The carrier gas was added about 85 cm upstream of the LMR detection region with a flow rate between 2 and 30 STP cm3/s, depending on the de- sired flow velocity and pressure. The flow rates were measured with mass flowmeters which are accurate within f3% [12]. The flow tube pressure was measured at the center of the reaction region with a capacitance ma- nometer with an accuracy of about fl% [12]. The flow tube pressure Pt ranged from 0.6 to 6.6 torr with the average flow vdocity B ranging from 350 to lo00 cm/s for the C10 + NO2 + M kinetic studies, while Pt = 1-3 torr and E = 1670-2730 cm/s for the C10 + NO studies.

Temperature variability of the flow tube from 202 to 416 K was achieved by circulating either liquid NZ, cooled ethanol, or heated dibutylphthalate through the flow tube jacket. The temperature of the circulating fluid was measured with a Pt resistance thermometer at the exit of the jacket, and it was shown previously that the temperature of the flow tube was uniform

The He (>99.9996%, analyzed) was passed through a liquid N2-cooled molecular sieve trap before use. 02 and N2 (>99.97%) and C12 (>99.96%) were used without further puriilcation. NO reactant was passed through a dry-ice-ethanol-cooled silica gel trap to remove other nitrogen oxides. NO2 was synthesized by reacting purified NO (>99.0%) with excess of 02, and stored under about 2 atm of 02 for a t least 24 h before use. During the kinetic measurements, the pure NO2 (NO < 0.01%) was kept in an ice-water bath to maintain a constant pressure of -260 torr.

0 3 was prepared by passing 0 2 through an ac discharge and collected in a silica gel trap kept in a dry-ice-ethanol bath. The ozone partial pressure Po3 was measured by passing He-03 mixtures through a 10-cm absorption cell using 1.15 X cm2 as the 0 3 absorption cross section at 254 nm [16]. Thus the flow rate of 03, Fo3, was calculated from the flow rate of He through the cell F H ~ , the total pressure in the cell PA, and P03:

to floc [12].

- (8) FOB = FH203/(PA .- Po,) .

The estimated random error in a rate constant measurement is about 8%. This is based on the equation k a F$T3 (sIope)/P;FRr2 and the esti- mates of 3% error in flow rate measurements FT and FR, 1% in temperature T, 2% in the decay plot slope, 1% in pressure Pt , and 1% in the tube radius r . However, in consideration of possible systematic errors, f15% is a more realistic estimate of the overall accuracy of the kinetic measurements.

- -

C10 Spectroscopy

The LMR detection of a paramagnetic radical requires a close coinci- dence of an FIR laser frequency with a zero-field rotational transition, because the energy range over which a molecular rotational transition can be magnetically tuned is usually less than 1 cm-1 for magnetic fields of less

*

716 LEE ET A&.

than 20 kG (104 G = 1 T). In the case of C10, the search for proper FIR laser frequencies that can yield LMR spectra is simplified because the molecular parameters for the ground state have been accurately determined [17-231. Using the energy expressionaj = B(J + W2 - D ( J + 1 / ~ ) ~ from the analysis of Kakar et al. [17] and the values of B = 0,61976 cm-l, D = 1.315 X lo6 cm-I for 35c10, and B = 0.60930 cm”, D = 1.278 X cm-l for 3’ClO, the zero-field rotational transition energies of C10 can be esti- mated with sufficient accuracy to predict near resonances. Listed in Table I are the estimated energies for the A J = +1 rotational transitions of the ground 2113/2 (u = 0) state of W l O and 37C10.

The magnetic tunability of a rotational transition can be estimated from the first-order Zeeman interaction energy

where gJ is the rotational g factor, p~ is the Bohr magneton, MJ is the component of J in the field direction, and H is the magnetic field strength. The gJ factor for C10 may be approximated by 1241

(A + Z)(A + 2Z) J ( J + 1) gJ 1

where J is the total angular momentum quantum number, A is the orbital angular momentum projection quantum number, and I: is the spin angular momentum projection quantum number. Thus of the two components

TABLE I. Estimated energies and vacuum wavelengths for AJ = +1 rotational transitions of the ground 2&/2 (u = 0) state of W I O and W10.

37c10 3 5 ~ ~ 3’ (lover level) E (an-’) Ulj”) E (m”) X(m)

.. 1.5 3.066 3283 3.099 3227 2 .5 6.265 2345 4.338 2305 3.5 5.483 1824 5.577 1793 6.5 6.701 1492 6.816 14 67 5.5 7.919 1263 8.055 124 1 6.5 9.137 1094 9.294 1076 7.5 10.35 965.7 10.53 949.6 8 .5 l.l.57 866.1 11.77 849.5 - 9.5 12.79 781.9 13.01 768.7

10.5 14.01 714.0 14.25 701.9 u. 5 15.22 656.9 15.40 665.0 u . 5 16.44 608.3 16.72 598.1 13 .5 17.65 566.4 17.96 556.9 16.5 10.87 530.0 19-19 521.0 l 5 . 5 20.08 . 497.9 20.43 409.5 16.5 21.30 469.5 21.66 461.6 17.5 22.51 446.2 22.90 436.7 10.5 23.72 421.5 24.13 414.4 19.5 24.92 401.0 25.36 396.2 20.5 26.15 382.6 26.60 376.0

- . .... . . .. . . . . .

LMR STUDIES~OF cio 717

of the *II C10 ground electronic state, the 211.1/2 levels are not &man active (because A = fl, 2 = ~1/2), while the 2113/2 (A = fl, u = f1/2) levels are Zeeman active. In addition, from eq. (10) one sees that the lower J tran- sitions exhibit a greater range of Zeeman tuning, and the requirement for a close match with FIR laser frequency is therefore less stringent. Un- fortunately the longest wavelength laser emission attainable in our spec- trometer was about 900 pm, limiting the attempt of C10 detection to the J = 9.5 - 8.5 and higher transitions.

The LMR spectra of W O and 37C10 have been observed at five different FIR laser frequencies as shown in Figures 1 and 2. Identical spectra were recorded by reacting C1 with C120 or 03, or by reacting 0 atoms with a large excess of C12. Thus the identity of the absorbing species is certainly ClO. Table II lists all the C10 rotational transitions detected, the FIR laser wavelengths, laser gases, and the laser electric vector polarization with respect to the magnetic field (K = parallel, u = perpendicular). The as- signments were made based on the molecular constants reported previously

For the J' = 11.5 + J" = 10.5 r-polarization transition of 37C10 (2113/29

u = 0), a detailed theoretical calculation was carried out2 The calculation involves diagonalization of an effective Hamiltonian matrix (3f,ff) between basis-set wave functions to obtain rotational energies as a function of magnetic field strength. In this study only the *II3/2 wave functions are

K171.

024 i s ' " I6 " 21 '

- % ~ @ ~ b @ ~ J

0 S I 4 17 21 Field (ki!opuss)

Figure 1. LMR spectra of C10. A--r spectrum of 35C10 (J = 14.5 - 13.5) using 556.9 pm CDJ line; B--a spectrum of aC10 (J = 11.5 - 10.5) using 699.5 pm CH30H line: C-same as A, except in (I polarization; D--a spectrum of 37C10 (J '= 16.5 - 17.5) using 469.1 pm CH3OH line.

2 For a complete derivation see [25].

718 LEE ET AI:

C . . . ) . . . . . . . 1 . . . . . . - . , 6 D 6 a n I S P

F& kilopuss)

Figure 2. LMR spectra of "C10. A-u spectrum of J = 10.5 - 11.5 transition using 713.1 pm DCOOH line; E&* spectrum of J = 17.5 - 18.5 transition using 444.4 pm CD3I line; C-same as B, except in u polarization; D-same as A, except in K polarization.

included in the basis set because C10 is a very good Hund's case (a) type molecule, and the mixing between the 2&/2 and 2111/2 levels is not impor- tant (A cv 319 cm-1). The C1 nucleus has a spin of I = 3/2, and since hy- perfine effects produce observable but not completely resolved splittings, nuclear spin effects are included. Thus the total angular momentum F = J + I and the basis set wave functions are characterized by the following quantum numbers:

01) \k = 1AS;slm) I @J'MF)

where S is the total spin angular momentum quantum number; m is the projection of I on the molecular axis and may have four values, 3/2,1/2,

TABLE 11. Summary of observed C10 spectra. ..

Detected j' + J" kser ~ . s \anex (m) k.er Line Polarbation Spec ie s CO, -

11.*10.5

17. S16.5

14.543.5

14.54-13.5

18. M 7 . 5

18.5+17.5

11.5ClO. 5

11.5+10.5

699.5

469.1

556.9

556.9

444.4

i44.4

713.1

713.1

9234

1OR38

10P36

10P36

9R32

9a32

low 10R34

- LMR STUDIES OF C10 719

-1/2, and -3/2; CP is the projection quantum number of F on the molecular axis; CP = A + Z + m; and MF is the projection quantum number of F in the direction of the magnetic field. The basis set consists of 32 wave functions corresponding to eight sequential values of F, and each F value generates four hyperfine sublevels corresponding to the four possible values of m.

The %,a is derived from a complete Hamiltonian, which is essentially the one given by Carrington [24]. Since the basis set contains only 'II312 wave functions, this Hamiltonian is simplified to a great extent, that is, all terms that mix the T I 3 1 2 ground state with states of different A or Z are neglected. The resulting %,E is given below, where it has been divided into four parts, spin rotation (sr), hyperfine (hf), quadrupole (Q), and Zeeman (Z):

(12) - A2 - X2 - m2 - F+I- - F J + ] + AAZ

where F* = F, f iFys A is the spin-orbit coupling constant, and B is the rotational constant.

(13) 3fM = amA + ( b + c)mZ where a, b , and c are Frosch and Foley hyperfine constants.

3f,, = B[F(F + 1) + S(S + 1) + I(I + 1) - CP2

e2Qq [3m2 - I ( I + l ) ] HQ = U(2z - 1)

where e*Qq is the quadrupole coupling constant.

(15) 3fz = PBHQZ~(A + gez) + grPNH(aZz + azxI.x + GzyIy) where azZ, azx, and azY aie direction cosine matrix elements, g, is the electron spin g factor, gr is the nuclear spin g factor, and P N is the nuclear magneton.

The matrix elements can be calculated exactly from known relations for the direction cosine matrix elements and ladder operators (see, for example, [26]). Nonzero matrix elements of Heff result only for diagonal elements and for those basis set functions that are connected by the quantum number changes A F = 0 or f l and/or A@ = Am = 0 or fl. A computer program which calculates these matrix elements as a function of F, MF, and H and - diagonalizes the resulting 32 X 32 matrix to yield rotational energies and - wave functions has been developed [25].

The program was used to simulab,the 37C10 a-polarized, J = 11.5 - 10.5 spectrum obtained by using the 713.1-pm DCOOH laser line. The ap- . propriate molecular constants for 37C10 are B = 0.609295 cm'l[17], A = -318.7 cm-l[19], a + 1/2(b + c) = 3.1305 X loe3 cm-l[21], e2Qq = 2.316 X 10-3 cm-1[17] and gr = 1.7395 X 10-8 cm-l.G-l[27]. For this transition the energ+ required for zero-field absorption is -14.009 cm-l (Table I). To achieve resonance with the FIR laser energy at -14.023 cm-l, this rotational

__

720 LEE ET AI=

transition must be &man tuned by an amount equal to 4.014 cm-1. The observed spectrum was simulated by calculating the magnetic fields re- quired for resonance with each hyperfine component of a J’, M; - J ” , Mi transition. Although the program is written in terms of the quantum numbers F and MF, the program output is easily interpreted in terms of J and MJ from the strong J dependence of the energies and from the relation MJ = MF - MI, where MI is the projection quantum number of I on the laboratory (magnetic field) axis. The results of the calculation are given in Table 111. The predicted magnetic field for each hyperfine component is also compared with the observed spectrum in Figure 3. The observed hyperfine structure is not well resolved. Only the transitions at the lower fields show structure on the high-field side of the absorption peaks. However, the observed contours are in qualitative agreement with the predicted hyperfine splittings, and the position of each hyperfine quartet is in excellent agreement with the observations. The relative in- tensities shown in Figure 3 are obtained by calculating a quantity propor- tional to the square of the dipole transition moment integral, (% I azz I

TABLE 111. Results of calculation for the J’ = 11.5 - J” = 10.5. s(Ah43 = 0) spectrum of 3’C10.

-10.5 8.398 8.43 8.370 8.412 8.510

-9.5 9.299 9.32 9.276 9.318 9.4l3 \

-8.5 10.420 10.44 1C.403 10.442 10.535

-7.5 l l . 8 5 5 l l . 8 7 11.844

11.971 1l.m

-6.5 13.757 13.73 13.752 13.792 13.875 .

-5.5 16.411 16.34 16.411 16.450 16.529

LMR STUDI&OF cio 721

Field (kilopaussl

Figure 3. Comparison of the calculated results with the observed r spectrum of 3TlO (J = 10.5 -c 11.5) using the 713.1-pm DCOOH laser line. A-observed; B-calculated.

91 ) 2, where the subscripts u and I refer to the upper and lower states, re- spectively. The relative intensity comparison between these calculated values and the LIvlR observations is only qualitative because other factors, such as field homogeneity and the absorption linewidth [28], affect the observed first derivative spectra. However, the overall agreement between the observed and calculated sp_ectra clearly indicates that we are detecting C10 in the X2&/2(u = 0, J = 11.5 - 10.5) transition.

Kinetic Results

Kinetic Studies of C10 + NO2 + M Reaction -

All the kinetic measurements were made under pseudo-first-order-con- ditions ([NO21 >> [ClO]) such that the NO2 concentration was constant during each measurement of the first-order C10 decay. Under such con- ditions, the rate expression for reaction (6) is written as

where iT is the average flow velocity, z is the distance from the end of the movable injector to the detection volume, [MI is the concentration of the third body, and k1, and third-order rate constants, respectively.

A set of typical C10 decay plots is shown in Figure 4. Each measurement was made by varying the reaction distance z by 20-30 cm. The decay of [ClO] in the measurements was greater than a factor of 1.5, but less than

and kru are th'e effective first-order, second-order, *

722 * *

LEE ET AL.

V : 4

I 1 I 10 20 30

Reaction Distance z (cm) Figure 4. Pseudo-first-order decay plot for C10 + NO2 + 0 2 reaction. T = 298 K, P, = 3.09 torr, D = 480 cm/s, [O,] = 8.46 X 10l6 molec/cm3, [He] = 1.44 X 1016 moledcm3. A-[N02] = 8.04 X 1014 molec/cm3; B-[N02] = 1.14 X 10'5 molec/ cm3; C [ N 0 2 ] = 1.50 X lOI5 moledcma. The X symbols are the C10 blank mea- surements showing negligible C10 wall loss, [NO21 = 0.

a factor of 15. The slope of the semilogarithmic plot multiplied by B gives the uncorrected first-order rate constant, kf, (in s-1). The rate constant was corrected for axial diffusion?

(17) k* = ki(1 + Dki/Gz) where D is the diffusion coefficient in cm2/s. The C10 diffusion coefficients were estimated as 490(T/300)1-7P-1 torr.cm2/s (P in torr) in He and 130 (T/300)1-7P-1 in N2 and 0 2 . The axial diffusion correction was small, generally less than 6%.

The X symbols near the top of Figure 4 indicate the C10 concentration measured with no NO2 added to the flow system. The near zero change in [CIO] at T > 250 K demonstrates that the C10 loss on the flow tube surface is negligible, k, < 0.1 s-1, and that there are no other reactions of c10.

Figure 5 shows a plot of kx versus [NO21 for the C10 + NO2 + 0 2 reaction. The results obtained from Figure 4, lines A-C, are marked A, By and Cy respectively. At each temperature [NO21 was varied while the rest of the experimental conditions remained unchanged. The excellent linearity and zero intercept of this plot show that the equation k' = kII [Nos] is valid,

The axial diffusion correction is given in [29a]; the diffusion coefficients are estimates based on data given in [29b].

-

-

.

[NO21 molecule ~ m - ~ ) Figure 5. Plot of k1 versus [NOJ for C10 + NO2 + 0 2 reaction. X-T = 298 K, Pt = 3.09 torr, slope = 2.04 X 10-1' cm3/molec-s; 0--T = 250 K, Pt = 1.24 torr, slope = 1.71 X 10-14 cm3/molec-s. [NO21 at this temperature has been corrected for dimerization. The results from Figure 4 are marked A, B, and C, respec- tively.

and the slope gives kI1. Some kxx measurements in this study were obtained by simply dividing the measured k* value by [NO&

A plot of klI versus [MI (total gas density) at 250 and 298 E; using He or 02 as a carrier gas is shown in Figure 6. The data shown in Fieme 5 for T = 298 and 250 K are indicated by A and B. The linearity of this plot demonstrates that the reaction is third order at pressures below 6 torr. -For

-

- c

IU)

fQ E 0 9 '0 Y

LMR STUDIES'OF CIO 723

[MI (10' molecule ~ m - ~ ) Figure 6. Plot of k" versus [MI for the C10 + NO2 + M reactions at 250 and 298 K. 0-M = 02; X-M = He. The data shown in Figure 5 are indicated by A and B.

724 LEE ET Ak.

experiments using He as the third body, the slope gives k E directly, because the concentrations of other species such as NO*, 0 2 , and C12 is negligible. However, for experiments using N2 or 02 as the third body the amount of He carrier added through the microwave discharge (ClO source) and the ozone reservoir comprises up to 30% of the gas mixture in the flow tube at low pressures. To derive accurate k# values for various third bodies, a linear least-square fit of the data to the expression

(18) kn = kfz + k;',[He] + k8:[02] + kR;[Nz] was carried out by computer. The intercept in the k" versus [MI plot, kz, is a pressure-independent heterogeneous contribution to k", which can be taken as the second-order rate constant for the C10 + NO2 reaction with the stabilization by wall collisions. Thus for a simultaneous fit, the kLf value is constrained to be the same for different carrier gases at the same tem- perature. The magnitude of k: is relatively small, (0.5-2.1) X cm3/molec.s, and varies with different wall conditions and temperatures. Therefore experimental data obtained at different times with different wall conditions were grouped separately, as shown in Table IV.

In Table IV, the experimental conditions, kg, and k# values for N2 at 297 K and for He and 0 2 at six temperatures ranging from 250 to 416 K are summarized. The listed uncertainties for k: and k)# are one standard deviation of the coefficients derived from the fits of the experimental data to eq. (18). At room temperature K # i was measured as (1.03 f 0.02), (1.06 f 0.02), and (1.07 f 0.02) X 10-31 cmVmolec2-s at three different times with

Z U b 6 1.10-5.92 1.10-5.W 0.31-2.- O2 I1 0.56-5.51 0.260.41 0.43-1.M l.5Sa.2

2% pr I8 0.864.88 0.1M.88 0.65-2.% 1.0?fo.O3

I 0.814-11 0.24-0.33 0.4W1.23 2.12M.03 4 14 0.164.11 0.254.38 0.13-1.61 l.ulto.03 1.9to.1

O.lfo.1 250 ma 9 0 .914 .46 0.914.46 0.384.13 2.2Ma.07-

O2 11 0 .114 .53 0.21-0.29 0.19-1.N 3.8biO.oS' - 272 Be I l . l M . 4 1 1.134.41 0.160.80

. O2 I 0.724.21 0.19-0.23 0 . 2 M . U l.5oiQ.M 1.7Jto.06 1.6iO.L

O . s M . 1 Y Z 8 b 9 0.95-4.46 0 . 9 5 4 . b b 0.80-1.72 0.7zto.m

O2 10 0.574.16 0.214.Y 0.92-1.69 l.lM.02

0.9to.2 Y l Ilt I 1.03-3.91 1.03-3.11 1.16-1.12 0.4Mo.03 0, 10 0.14-0.2) 0 . 2 k 4 . 2 9 1.48-1.71 0.76fo.OS

I16 O2 7 0.71-3.58 0.31-0.40 1.33-1.85 0.58to.03' O . M . 1

~

a Measured with correction for NO? dimerization. Without correction, k{L and k#

b Assuming k;', = 3.8 X cm6/molec2-s. See text. were 2.17 and 3.69 X cm6/molec2-s, respectively.* See text.

- LMR STUDIES OF C10 725

excellent consistency. Thus an average value of 1.05 X cm6/molec2.s is obtained for the C10 + NO2 + He reaction at room temperature. Simi- larly the room temperature values of 1.87 X 10-31 and 2.09 X 10-31 cm6/ molec2-s are obtained for M = 0 2 and M = N2, respectively.

At 416 K the C10 + NO2 + He reaction was too slow to be measured. The data for M = 02 a t this temperature were corrected for the small contri- bution from M = He using kfi: = 3.8 X cm6/molec2-s obtained by extrapolating the lower temperature data.

In Figure 7 the data for kg and kg are plotted as a function of UT. The least-square fits gave kE = (2.79 f 0.38) X exp[(1087 f 41)/T] and

error limits are one standard deviation limits on the coefficients. As mentioned earlier, the concentration of N204 becomes significant only

at the lowest temperatures. An attempt was made to measure the rate constant for the reaction of C10 with N204. A special double port movable inlet was constructed. One port was designed to maximize the N204 con- centration injected into the flow tube. It had a short ( N 1.5 cm) length of 0.09-mm inside diameter capillary attached to the end of a 14-mm inside diameter X 10-cm length bulb. This constriction caused a high pressure (-40 torr) to develop inside the inlet and therefore enhanced the NO2 di- merization. The pressure in the inlet was measured and the fractional dimerization was calculated using the equilibrium constant [15]. A t 255

k111 o2 = (3.48 f 0.22) X lo-= exp[C1183 f 19)/T] cm6/molec2-s, where the

T ( K ) 500 400 300 250 200 IOL 1 I I I

1 1

t 1 1 3 D 4D SO

I O O O I T ( K-’1

Figure 7. Arrhenius plot of k a and for C10 + NO2 + M reaction.

726 LEE ET k

K the N204 flow rate was typically 7040% of the total flow rate (N204 + NOa). Equilibrium was assured by the long residence time in the cooled inlet tube (-7 s). The second port was designed for minimum dimerization, with low pressure and a short residence time. For these measurements the C10 was added to the carrier gas upstream from the reaction zone. When a mixture of about 75% N204 and 25% NO2 was reacted with C10, the ob- served C10 loss rate was significantly Iess (-1/2) than when the same flow rate was added through the second injector port as monomers. From this observation we conclude that small amounts of N204 do not interfere with the C10 + NO2 kinetic measurements. The N204 is probably not disso- ciated significantly in the flow tube at 255 K and pressures of a few torr. If that is true, then k(C10 + Ni04) N 2 X cm3/molec-s.

Kinetic Studies of ClO + h'0 - Cl + NO2 Reaction

A typical C10 decay plot is shown in Figure 8. In these measurements the reaction distance z was varied by 15-30 cm. The C10 decay with no NO added (X symbols at the top of Figure 8) also indicated negligible wall loss. The axial diffusion correction was applied and was less than 8% at 209 K and 3% at 393 K.

Figure 9 shows a plot of kl as a function of [NO] for data at 393,296, and 202 K. At room temperature, some measurements (indicated A on Figure

IO . 20 30 t I . , : 1

'0 Reoction Distonce L (cm)

Figure 8. Pseudo-first-order decay plot for C10 + NO reaction. T = 393 K, Pt = 1.91 torr, i7 = 2210 cm/s. A-[NO] = 5.27 X lor2 molec/cm3; B-[NO] = 1.19 x 1013 molec/cm3; C-[NO] = 2.01 X 1013 molec/cm3. The X symbols are the C10 measurements showing negligible C10 wall loss, [NO] = 0.

LMR STUDIES OF C10 727

700 i

[NO] (IO’’ moleark ~ 6 ~ )

Figure 9. Plot of k1 versus [NO] at 202,296, and 393 K for C10 + NO reaction. At 296 K. +-NO through the movable injector; A-CIO through the movable injector. A, B, and C indicate data from Figure 8.

9) were performed by adding C10 through the movable inlet and adding NO with the w r i e r gas. No difference in rate constant was observed. No pressure dependence was observed over the range of 1-3 torr. The data for k1 versus [NO] a t each temperature were fitted with a least-squares program to a straight line. The slope kI1, intercept, and experimental conditions at five temperatures ranging from 202 to 393 K are summarized in Table V. At room temperature the rate constants were measured as 1.66, 1.84, and 1.84 X 10-11 cm3/molec.s at three different times. These give - an average value of 1.78 X cm3/molec-s. .-

I I 3.0 4.0 50

IOOO/T ( K-’1

Figure 10. data.

Arrhenius plot for C10 + NO - C1 + NO2 reaction rate constant

728 - .

LEE ET AL.

195 10 O.KZ.56 0.13-2.67

zn 17 1.0S-J.10 0.81-2.91

291 s 1.11 0.41-1.10

1m 10 1.06-1.93 0.6w2.19

217 9 1.01-1.w 0.U-2.60

1.uLo.01 otr

1.Llia.D) 1 W

1 . u t o . 0 1 1ois

2.UtO.06 14i4

z.sz+O. os 1&7

23b 9 1.W-1.95 0.63-2.10 2.2w.w 7 H

1Y 10 1.11-1.u 0.92-5.19

131 10 1.01-2. U 0.M-2.77

391 10 1.914.77 0.53-3.10

2 . i w . m -1lil2

1.61tO.m -16

1.JW.01 -w

The measured rate constants are plotted versus 1/T in Figure 10. A least-squares fit to the Arrhenius expression gives k3 = (7.13 f 0.48) X lo-’* erp [(271 f 18)/T] cm3/molec.s, where the error limits represent one standard deviation.

Discussion

T h e measurements of k (C10 + NO:! + M) are summarized with the work from other laboratories in Table VI. The error limits have been increased to allow for systematic errors. Our M = He measurement is about in the middle of the values obtained by Leu et al. [8] and Zahniser et al. [lo] and agrees well with the 297 K value reported by Birks et al. [9]. The range of these measurements at room temperature is about f15%. For M = N2 we measured k only at 297 K. Our result is higher than all of the published values. The maximum discrepancy is 38% compared to Zahniser et al. Although the error bars on these measurements nearly overlap, this dif-

TABLE VI. Comparison of results for C10 + NO2 + M reaction.

k(ZS¶O M Arrhcni”. reon (ld’1c.6DIecul. -’*-’) (t:rr) T . c b l p u r . bfer-a

250-387 250-365 148-517

297

1 . o w . 1 5 0 . l l M . U

0.*6M.01 1.2qo.m

0.67-6.60 1.1 -6.s

1-9 2.5 -5.0

0.61-1.70 2 4 . 1 1-6

1 . 5 4 . 5 50 20

0.56-5.57

1-4

297 2.opto.w a97 1.529.23

2911417 1.72H.17 15D-lSb l . ls10.05 276-339 1. u

1.5 H.12

210-411 l . l W . 2 8

29a 1.im.10

a DF-discharge flow; LMR-laser magnetic resonance; .RF-resonance fluorescence; MS-mass spectrometry; MA-modulated absorption; FP-A-flash photolysis ultraviolet absorption.

LMR STUDIESaF C10 729

ference is larger than is normally found. This work is the only reported measurement with M = 0 2 . We can compare this result with the earlier work on the basis of the temperature dependence, which agrees very well with previous measurements in He and N2.

Cox and Lewis (301 have reported the most extensive pressure depen- dence study of reaction (6). They measured the rate constant in N2 over the range of 25-612 torr using modulated photolysis with ultraviolet ab- sorption detection of C10. Their data connect reasonably well with the

sures greater than 50 torr. Molina et al. [34] have measured k6 at 298 K in 20 and I00 torr N2 using

a flash-photolysis ultraviolet absorption technique. Their result using a ClzO source of C10 agrees reasonably well with the others.

The product of the termolecular reaction of C10 and NO2 was not quantitatively observed here or in most of the previous studies [%-lo]. It has been suggested [31-341 that more than one isomer of ClN03 may be formed. The possibilities include the following

1 low-pressure data, as seen in Table VI. They report some falloff at pres-

( 6 4 C10 + NO;! + M --c ClOKOz + M (6b) - ClOONO + M (6c) - OClNOp + M (6d) - OClONO + M

The studies of the thermal decomposition of CION02 [31] and the isotopic exchange reaction between ClON02 and 'jNO;! by Knauth and co-workers [32], when combined with the equilibrium constant for ClONOz formation, lead to a value of kfli that is about 30% of the values listed in Table VI. I t is not clear that this difference is significant due to uncertainties in the thermochemical data. Chang et al. [33] suggested, on the basis of a re- combination theory calculation, that if the isomer ClOONO is formed along with ClONO;!, the discrepancy between the forward and reverse rate con- stants of ClN03 can be resolved. Recently, in a flash-photolysis experi- ment, Molina et al. [34] have observed different rates of removal of ClO depending upon the C10 precursor ((2120 or OC10). They proposed that an OClONO isomer is formed and takes part in a secondary reaction. Using infrared absorption spectroscopy, they also found that the amount of CION02 produced was only about one third of the amoun~ oh C10 removed. Thus it is probable that more than one product is formed in this reaction. This conclusion has a very important implication to atmospheric chemistry because the iaentity of the product molecule determines the photolysis rate:, and products and hence the effect of the C10 + NO;! + M reaction. The large rate constant for the reaction makes it a potentially important proce,t;s for C10 in the stratosphere. Therefore the reaction products must be de- termined and evaluated.

730 LEE ET

The measurement using 0 2 as a carrier gas was carried out to investigate the possib1it.y of a reaction rate enhancement by 02. The reaction of 0 2 with C10 to f’orm OC102,

(7) C10 + 0 2 + M i= OC102 + M may be possible,4 although no information on the reaction rate is available, If reaction (7) .is sufficiently fast and followed by the reaction

(19) 0C102 + NO2 - ClN03 + O2 the rate of ClNt03 formation could be enhanced in oxygen. Assuming a steady state of [ClO3], the rate constant enhancement is k19k7[02]l(k-7[M] + k19[N02]), where k-7 is the back-reaction rate constant for reaction (7). In the present experiment the variation of [02] and [NO,] was not large enough to observe a significant enhancement due to this mechanism.

The measured relative third-body efficiencies of 1.01.82.0 for He, 02, and N2, respectively, in C10 + NO2 + M reaction can be compared with those obtained fox the similar reaction

J

(20) Hop + NO2 + M --L HOON02 + M by Howard [36], namely, 1.&1.5:2.1. Although the 0 2 third-body efficiency measured in this study for reaction (6) is slightly higher than that of reaction (20), the difference is too small to be considered significant.

The activation energy of kg:, (1180 f 80), is slightly higher than that of kpe, (1090 f 80). Although the difference may be due to the experimental uncertainties, the possibility remains that this is an indication of a different mechanism for C10 + NO2 reaction in 0 2 . We plan to extend these mea- surements to higher pressures as a further test for a mechanism involving OClO*.

The measurements of the C10 + NO rate constant k3 are summarized

TABLE W. Comparison of rate constant measurements for C10 + NO - c1+ NO2 reaction.

l(12-393 (I .1*1.4 ).lo-~.yl(27am)/Tl 1.7w.n D I - W I+& .art

2u 1.7 i0.1 D I 4 C l p . d U r c . a J - 130.295 (1.13tQ.l~)x10~11.~I(2DQt30)/T1’ 1.1 MA’ D I - U 2 . IP l~r d Lmuf-m4 - tll4lS (~.T~M.lE)~O~U.4i(2)~O)/T) 1.53s.11 DI- Lr rl Qkr.’

2 U l.%.l* ~1-16 tx,wrlmelob.rr*

I* 1.7W.15 M46 ..,.1(yItP’

\ a DF-discharge flow; LMR-laser magnetic resonance; MS-mass spectrometry; RF-

b Based on a rate constant ratio measurement as described in text.

4 Prasad [35] also recently discussed some possible implications of OC102 in stratospheric

. resonance fluorescence.

chemistry.

LhlR STUDIES OF C10 731

in Table VII. Our room temperature measurement is in excellent agree- ment with those of Clyne and Watson [3] and Ray and Watson [7] and in good agreement with the other measurements. The temperature depen- dence is very close to the result of Leu and DeMore [5], although their rate constants fall about 15% below ours. The temperature dependence given by Zahniser and Kaufman [4] is also in good agreement, but their values fall somewhat higher than ours. The k3 given in Table VI1 for Zahniser and Kaufman is based on their measurement of the ratio of k3 to k(C1-k 0 3 ) and their value for k(C1 + Os).

In conclusion, we fmd that the LMR technique can be used to detect C10 radicals with sufficient sensitivity allow pseudo-first-order kinetic studies in a flow-tube reactor. This method offers the advantages of positive spectroscopic identification of the C10 with no interference from product species or from problems encountered with indirect measurement methods.

Acknowledgment

This work was supported in part by the Chemical Manufacturers Asso- ciation Technical Panel on Fluorocarbon Research. The authors wish to thank J. T. Hougen for valuable advice on the interpretation of the LMR spectra.

Bibliography

111 F. S. Rowland and M. J. Molina, Rev. Geophys. Space Phys.. 13,1(1975). (21 As described later, several different isomers of ClNOa are possible. The formula ClN03

as it is used here is not meant to imply any particular isomer, only the empirical formula of the product of reaction (6).

[3] M. A. A. Clyne and R. T. Watson, J. Chem. Soc., Faraday Trans. I , 70,2250 (1974). [4] M. S. Zahniser and F. Kaufman, J. Chem. Phys., 66,3673 (1977). (51 M. T. Leu and W. €3. DeMore. J. Phys. Chem., 82,2049 (1978). [SI M. A. A. Clyne and A. J. MacRobert, Int. J. Chem. Kinet., 12,79 (1980). I71 G. W. Ray and R. T. Watson, 14th Informal Conf. on Photochemistry, Newport Beach,

[8] M. T. Leu, C. L. Lm, and W. B. DeMore, J. Phys. Chem., 81.190 (1977). [9] J. W. Birks, B. Shoemaker, T. J. Leck, R. A. Borders, and L. J. Hart, J. Chem. Phys.,

-.

_.

CA, 1980.

66,4591 (1977). - [lo] M. S. Zahniser. J. S. Chang, and F. Kaufman, J. Chem. Phys., 67,997 (1977). [11] C. J. Howard and K. M. Evenson, J. Chem. Phys., 61,1943 (1974). [12] R.,M. Stimpfle, R. A. Perry, and C. J. H0ward.J. Chem. Phys., 71,5183 (1979). [13] R.;T. Watson, J. Phys. Chem. Ref. Data, 6,871 (1977). [14] R. J. Nordstrom and W. H. Chan,J. Phys. Chem., 80,847 (1976). [15] D. R. Stull and H. Prophet, Eds., “JANAF Thermochemical Tables,” Nat. Bur. Stand.,

[16] U‘. B. DeMore and 0. Raper, J. Phys. Chem., 68,412 (1964). [17] R. K. Kakar, E. A. Cohen, and M. Geller, J. Mol. Spectrosc., 70,243 (1978). [IS] T. Amano and E. Hirota, J. Mol. Spectrosc., 66,185 (1977). [191 U. Uehara, M. Tanimoto, and Y. Morino, Mol. Phys., 22,799 (1971).

Washington, DC, including revised tables issued through 1979.

732 LEE ET AL.

[20) A. Carrington, P. N. Dyer, and D. H. Levy, J. Chem. Phys., 47,1756 (1967). [21] T. Amano, S. Saito, E. Hirota, Y. Morino, D. R. Johnson, and F. X Powell, J. Mol.

[22] J. A. Coxon, W. E. Jones, and E. G. Skolnik, Can. J. Phys., 54,1043 (1976). [23] R T. hlenzies, J. S. hlargolis, E. D. Hinkley, and R A. Toth, Appl. Opt., 16, 523

[24] A. Carrington, “Microwave Spectroscopy of Free Radicals,” Academic, New York, 1974,

[25] R. M. Stimpfle, P b D . dissertation, Department of Chemistry. University of Colorado,

[26] J. T. Hougen, “The Calculation of Rotational Energy Levels and Rotational Line In-

[Ti’] C. H. Tomes and k L Schawlow, “Microwave Spectroscopy,” McGraw-Hill, New York,

(281 J. T. Hougen, J. Mol. Spectrosc., 54,477 (1975). 1291 (a) F. Kaufman, B o g . React. Kinet., 1.3 (1961); (b) T. R Marrero and E. A. Mason. J.

[30] R. A. Cox and R. Lewis, J. Chem. SOC., Faraday Trans. I . 75,2649 (1979). [31] H. D. Knauth, Ber. Bunsenges. Phys. Chem.. 82,212 (1976). (321 G. ScMnle, H. D. Knauth, and R. N. Schindler, J. Phys. Chem., 83,3297 (1979). [33] J. S. Chang, A. C. Baldwin, and D. M. Golden,J. Chem. Phys., 71,2021 (1979). (341 M. J. Molina, L. T. Molina, and T. Ishiwata, J . Phys. Chem., 84,3100 (1980). [35] S. S. Prasad, Nature, 265,152 (1980). [36] C. J. Howard, J. Chem. Phys., 67,5258 (1977).

Specrosc., 30,275 (1969).

(1977).

p. 192. _ _

Boulder, CO, 1979, chap. 2

tensities in Diatomic hlolecules,” NBS Monograph 115, June 1970, p. 31.

1955, p. 644. 4

Phys. Chem. Ref. Data, 1,3 (1972).

Received October 14,1981 Accepted January 4,1981